A real-time task scheduling method for cloud-edge collaboration in intelligent connected vehicles

Abstract

The present invention discloses a cloud-edge collaborative real-time task scheduling method for intelligent connected vehicles, including collecting real-time status data, task requirements and network conditions of the vehicle, performing multi-source data fusion processing, and generating comprehensive data; based on the comprehensive data, adjusting the task priority using a dynamic task priority optimization mechanism, and selecting the optimal path for task transmission using a shortest path algorithm; allocating tasks through collaboration between edge nodes to achieve dynamic migration and load balancing of tasks, identifying the scene type around the vehicle based on semantic analysis technology, and dynamically adjusting the task scheduling strategy; executing tasks according to the optimized scheduling plan, and monitoring the task execution status in real time, and updating the scheduling logic. The present invention improves the task scheduling efficiency, reliability and energy consumption optimization capabilities of intelligent connected vehicles as a whole, and provides strong technical support for achieving safe and efficient operation of intelligent connected vehicles.

Description

Technical Field

[0001] The present invention relates to the field of intelligent connected vehicle management technology, and in particular to a cloud-edge collaborative real-time task scheduling method for intelligent connected vehicles. Background Art

[0002] With the rapid development of intelligent connected vehicle (ICV) technology, vehicles are required to handle a large number of complex tasks during operation, such as path planning, environmental perception, and communication interaction. These tasks place higher demands on real-time performance, reliability, and resource optimization. Current mainstream task scheduling methods primarily rely on a single architecture for cloud computing or edge computing, making it difficult to achieve efficient task allocation in cloud-edge collaborative scenarios. Furthermore, existing methods have limitations in multi-source data fusion, dynamic adjustment of task priorities, network transmission path optimization, and edge node collaboration, making them unable to fully address the complex and ever-changing operating environment of ICVs. Furthermore, with the increasingly dynamic nature of the scene surrounding the vehicle, how to quickly adapt to the characteristics of scenarios such as traffic accidents, obstacles, and construction areas, and how to accurately optimize scheduling based on the vehicle's energy consumption, have become key bottlenecks hindering the further development of ICV technology. Summary of the Invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides a real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles, which specifically solves the following technical problems: First, how to efficiently integrate vehicle sensor data, V2X communication data and task history records to generate comprehensive task scheduling decision information; second, how to dynamically adjust task priorities based on task delay tolerance, resource requirements and dependencies to improve the flexibility of scheduling schemes; third, how to combine network topology information and real-time network conditions to optimize transmission paths under multi-objective constraints of delay, reliability and energy consumption; fourth, how to achieve dynamic task migration and resource load balancing through the collaboration of edge nodes to improve the utilization efficiency of edge resources; fifth, how to optimize task allocation strategies based on the remaining power of the vehicle and the energy consumption requirements of the task to ensure reasonable distribution of energy consumption; sixth, how to use semantic analysis technology to identify complex environmental scenarios, and dynamically adjust task scheduling strategies based on scenario types, so that the system can quickly adapt to and operate efficiently in a changing environment.

[0005] To solve the above technical problems, the present invention provides the following technical solution: a real-time task scheduling method for cloud-edge collaboration in intelligent connected vehicles, comprising:

[0006] Collect real-time vehicle status data, mission requirements, and network conditions, and perform multi-source data fusion processing on vehicle sensors, V2X communications, and historical mission records to generate comprehensive data;

[0007] Based on comprehensive data, the dynamic task priority optimization mechanism is used to adjust task priorities, and the shortest path algorithm is used to select the optimal path for task transmission based on network topology information, real-time network status, and energy consumption status;

[0008] By allocating tasks collaboratively between edge nodes, dynamic task migration and load balancing are achieved. Task allocation and resource allocation strategies are dynamically adjusted based on the vehicle's remaining battery life and task energy consumption requirements. Semantic analysis technology is used to identify the scene type around the vehicle and dynamically adjust the task scheduling strategy.

[0009] Execute tasks according to the optimized scheduling plan, monitor task execution status in real time, and update scheduling logic.

[0010] As an optimal solution for the real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles described in the present invention, the multi-source data fusion processing includes extracting time series and spatial features of sensor data, V2X communication data and task record data through a deep learning model. The time series features are extracted by a bidirectional long short-term memory network, and the spatial features are constructed by a graph neural network combined with an attention mechanism to construct a dynamic relationship between data nodes to generate a comprehensive data set for scheduling decisions.

[0011] As an optimal solution of the real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles described in the present invention, the dynamic task priority optimization mechanism includes dynamically adjusting the task priority using a reinforcement learning algorithm, adjusting the task priority weight based on the delay tolerance, resource requirements and dependencies of the task by defining a state set, an action set and a reward function, and updating the task scheduling order according to real-time changes.

[0012] As a preferred solution of the real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles described in the present invention, the method of selecting the optimal path for task transmission using the shortest path algorithm includes calculating the path weight based on network topology information and real-time network status, combined with delay, reliability and energy consumption constraints, and selecting the task transmission path using the shortest path algorithm, while verifying whether the delay, reliability and energy consumption constraints of the path meet the scheduling requirements. The weight formula is:

[0013] w ij =α·d ij -β·ln(r ij )+γ·c ij

[0014] Among them, w ij Represents edge e ij The comprehensive weight of α, β, and γ represent weight coefficients, and d ij represents the communication delay between nodes i and j, r ijrepresents the communication reliability between nodes i and j, c ij represents the communication energy consumption between nodes i and j.

[0015] As a preferred solution of the cloud-edge collaborative real-time task scheduling method for intelligent connected vehicles described in the present invention, the collaborative allocation of tasks between edge nodes includes constructing a task dependency graph to represent the execution relationship and order between tasks, and calculating the priority formula:

[0016]

[0017] Determine the priority of task scheduling and dynamically migrate tasks to nodes with loads below a preset threshold based on the load status and resource availability of edge nodes;

[0018] Among them, T i represents the task, P(T i ) represents task T i priority, α1, β1, γ1 represent weight coefficients, T delay Denotes the delay tolerance of task Ti, D dep Represents task T i The dependency depth, R res Represents task T i The amount of resources required, R avail Indicates the current available resources of the task execution node.

[0019] As a preferred solution of the real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles described in the present invention, the dynamic adjustment of task allocation and resource allocation strategy includes prioritizing low-energy consumption tasks by calculating the current remaining power of the vehicle and the energy consumption requirements of the task. The priority calculation formula is:

[0020]

[0021] And adjust the computing resource allocation capacity of the edge node according to the remaining power status of the vehicle, and dynamically allocate it to R alloc Multi-level allocation strategy;

[0022] Among them, P task Indicates the energy consumption priority of the task, E remain Indicates the remaining power of the vehicle, E task Indicates the energy consumption requirement of the task.

[0023] As a preferred solution of the real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles described in the present invention, the semantic analysis technology includes using a neural network model to perform semantic segmentation on the vehicle's surrounding environment data. The model uses an attention mechanism to improve the recognition ability of scene types such as traffic accidents, obstacles or construction areas, and adjusts the task priority based on the scene type. The adjustment formula is:

[0024] P task,i =P base,i +δ·R(l i )

[0025] Among them, l i is the scene label, R(l i ) is scene l i The weight adjustment value for task priority, P task,i represents the adjusted priority of task i, P base,i represents the basic priority of task i, and δ represents the scenario impact factor.

[0026] As a preferred solution of the real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles described in the present invention, the shortest path algorithm selects the optimal path for task transmission, including caching the data required for critical tasks at the edge node, reducing the access to cloud resources through cache management, and selecting cached data based on the priority, access frequency and dependency of the task.

[0027] A computer device includes a memory and a processor, wherein the memory stores a computer program, and is characterized in that when the processor executes the computer program, it implements the steps of a real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles.

[0028] A computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the computer program implements the steps of a real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles.

[0029] Beneficial effects of the present invention: The present invention achieves technical breakthroughs in data fusion, task priority adjustment, path optimization, edge collaboration, energy consumption-aware scheduling, and dynamic adjustment of semantic analysis by constructing a real-time task scheduling method for a cloud-edge collaborative architecture, and has the following beneficial effects: First, it realizes the efficient fusion of multi-source data, extracts temporal and spatial features through deep learning, and provides accurate data support for scheduling decisions; second, it dynamically adjusts task priorities through reinforcement learning, so that the scheduling scheme can adapt to real-time changing task characteristics and environmental conditions; third, it optimizes the transmission path in combination with network topology and real-time network conditions to meet delay, reliability, and energy consumption constraints; fourth, it balances computing loads and improves resource allocation efficiency through task collaboration and dynamic migration between edge nodes; fifth, it optimizes task allocation strategies through energy consumption-aware models, prolongs vehicle operating time, and improves energy utilization efficiency; sixth, it uses semantic analysis technology to identify scene types such as traffic accidents and obstacles in real time, and dynamically adjusts scheduling strategies to enhance the system's adaptability to complex environments. In summary, the present invention improves the task scheduling efficiency, reliability, and energy consumption optimization capabilities of intelligent connected vehicles as a whole, and provides strong technical support for achieving safe and efficient operation of intelligent connected vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0031] Figure 1 A flowchart of a real-time task scheduling method for cloud-edge collaboration in intelligent connected vehicles is provided as an embodiment of the present invention. DETAILED DESCRIPTION

[0032] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0033] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0034] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0035] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0036] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0037] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0038] Example 1, reference Figure 1 , which is the first embodiment of the present invention, provides a real-time task scheduling method for cloud-edge collaboration of intelligent connected vehicles, including:

[0039] S1: Collects real-time vehicle status data, mission requirements, and network conditions, and performs multi-source data fusion processing on data from vehicle sensors, V2X communications, and historical mission records to generate comprehensive data.

[0040] S2: Based on comprehensive data, the dynamic task priority optimization mechanism is used to adjust the task priority, and the shortest path algorithm is used to select the optimal path for task transmission based on network topology information, real-time network status and energy consumption status.

[0041] S3: Dynamic task migration and load balancing are achieved through collaborative task allocation between edge nodes. Task allocation and resource allocation strategies are dynamically adjusted based on the vehicle's remaining battery power and task energy consumption requirements. Semantic analysis technology is used to identify the scene types around the vehicle and dynamically adjust task scheduling strategies.

[0042] S4: Execute tasks according to the optimized scheduling plan, monitor the task execution status in real time, and update the scheduling logic.

[0043] The multi-source data fusion processing includes extracting temporal and spatial features from sensor data, V2X communication data, and mission record data through a deep learning model. Temporal features are extracted by a bidirectional long short-term memory network, and spatial features are extracted by constructing dynamic relationships between data nodes through a graph neural network combined with an attention mechanism to generate a comprehensive data set for scheduling decisions.

[0044] Specifically, the vehicle operating status data S is collected from the sensor t ={s t,1 ,s t,2 ,…,s t,n}, such as speed, position, acceleration, and environmental obstacle positions.

[0045] Obtaining traffic infrastructure data from V2X communications t ={v t,1 ,v t,2 ,…,v t,m}, including road traffic conditions and nearby vehicle data.

[0046] Extract data H from historical task records t ={h t,1 ,h t,2 ,…,h t,k}, including task type, resource consumption and network conditions.

[0047] Normalize the data and express it as X t =[S t ,V t ,H t ].

[0048] The improved bidirectional long short-term memory network (Bi-LSTM) model is used to capture the temporal dynamic relationship of the data. The formula is as follows:

[0049]

[0050] in, and are the forward and backward hidden layer states, W, U, b are the learnable parameter matrices, σ is the activation function, W f 、W bRepresents the weight matrix of the forward and backward LSTM, U f 、U b represents the weight matrix of the forward and backward LSTM, b f 、b b Represents the bias terms of the forward and backward LSTM.

[0051] The final output timing characteristics are:

[0052]

[0053] We use an attention mechanism combined with a graph neural network (GNN) to model the spatial relationships between different data sources. We construct a graph G = (V, E), where V represents the set of data nodes and E represents the correlation between nodes (such as physical proximity and task association). We extract features through the following steps:

[0054]

[0055] Among them, W Q ,W K ,W V is the learnable weight matrix, is the neighbor set of node i, X i 、X j Represents the input feature vectors of node i and node j respectively, d k represents the dimension of the key vector, X′ i represents the updated feature representation of node i.

[0056] The time series feature H t and spatial features X′ i Fusion is performed to generate comprehensive features F t =[H t ,X′ i ].

[0057] The dynamic task priority optimization mechanism includes dynamically adjusting task priorities using a reinforcement learning algorithm, adjusting task priority weights based on the delay tolerance, resource requirements, and dependencies of the tasks by defining a state set, an action set, and a reward function, and updating the task scheduling order based on real-time changes.

[0058] Specifically, build a reinforcement learning environment and define the state S t The current task information of the vehicle, including the task delay tolerance, resource requirements, task dependencies and vehicle operation status (such as remaining computing resources, network conditions). The state is represented by S t ={s t,1 ,s t,2 ,…,s t,n}, where s t,iRepresents the characteristics of the i-th task.

[0059] Define the action set A of reinforcement learning t , that is, the operation of adjusting the priority of tasks in the current state, such as adjusting the order of tasks or redistributing priority weights. The action set is A t ={a t,1 ,a t,2 ,…,a t,m}.

[0060] Design reward function R t , which is used to evaluate the effectiveness of the scheduling strategy after priority adjustment. The reward function comprehensively considers the timeliness of task completion, the rationality of resource allocation, and the degree of satisfaction of task dependencies. The specific formula is:

[0061]

[0062] Among them, T delay is the task completion delay, R res is the resource allocation rate, R max is the maximum available resource, D dep is the satisfaction of task dependency, and α, β, γ are weight coefficients.

[0063] The priority optimization strategy is trained using the Deep Q Learning (DQN) algorithm. The Q value function is defined as:

[0064]

[0065] Among them, λ is the discount factor, which represents the trade-off between current rewards and future rewards.

[0066] The optimal action output by the reinforcement learning algorithm Dynamically adjust the priority weight W of the task t The priority update formula is:

[0067] W t,i =W t-1,i +ΔW t,i

[0068] Where ΔW t,i is the priority adjustment amount calculated based on the Q value, W t,i is the priority weight of the i-th task at time t, W t-1,i is the priority weight of the i-th task at time t-1.

[0069] According to the updated task priority weights, the tasks are sorted and the scheduling order is generated. t ={o t,1 ,o t,2 ,…,ot,k}, where o t,i Represents the task with priority i.

[0070] Through this optimization mechanism, the reinforcement learning algorithm can dynamically adjust the priority ranking of tasks according to real-time changes in the environment, generate reasonable scheduling plans, and ensure the satisfaction of task dependencies and the rational allocation of resources.

[0071] The method of selecting the optimal path for task transmission by using the shortest path algorithm includes constructing a communication network topology structure of an intelligent connected vehicle and defining a node set V in the network = {v1, v2, ..., v n} and edge set E = {e ij ∣v i ,v j ∈V}, where each edge e ij Represents the communication connection between two nodes. Each edge is assigned relevant parameters, including the communication delay d ij , reliability ij (i.e., connection success rate) and transmission energy consumption c ij .

[0072] Use the real-time monitoring system to collect the current status data of each connection in the network, including bandwidth utilization, packet loss rate and transmission stability, and dynamically update the data. ij ,r ij ,c ij The value of .

[0073] Calculate the comprehensive weight value w for each edge ij , considering the constraints of delay, reliability and energy consumption, the formula is:

[0074] w ij =μ·d ij -π·ln(r ij )+σ·c ij

[0075] Among them, w ij Represents edge e ij The comprehensive weight of μ, π, σ represents the weight coefficient, d ij represents the communication delay between nodes i and j, r ij represents the communication reliability between nodes i and j, c ij represents the communication energy consumption between nodes i and j.

[0076] Use an improved shortest path algorithm (such as Dijkstra or A*) to search for the node v from which the task originates. s To the target node v t The optimal path. The algorithm steps are as follows:

[0077] Initialize the path costs of all nodes to infinity, starting from the node v s The path cost is set to 0.

[0078] Traverse the neighbor nodes of the starting node, calculate the path cost from the starting point to the neighbor node, and update the path cost of the neighbor node.

[0079] Repeat the above steps until the target node v t The final optimal path is the set of links with the minimum path cost.

[0080] Perform constraint verification on the searched optimal path to ensure that the path meets the following conditions:

[0081] Delay constraint: total delay D of the path p =∑ (i,j)∈p d ij Less than the maximum delay tolerance of the task.

[0082] Reliability constraint: total reliability R of the path p =∏ (i,j)∈p r ij Above the minimum reliability requirements.

[0083] Energy consumption constraint: total energy consumption C of the path p =∑ (i,j)∈p c ij Less than the energy budget that can be allocated to the system.

[0084] Here, p represents a path.

[0085] If the optimal path found passes the constraint verification, the path is selected for task transmission. If not, the path weight coefficients μ, π, and σ are dynamically adjusted and the path search is repeated. After the task transmission is completed, the real-time network status parameters of the path are updated.

[0086] The collaborative task allocation between edge nodes includes: n}, a directed acyclic graph is constructed based on the dependency relationship of the tasks. In the task dependency graph G = (V, E), the node V represents the task and the edge E represents the dependency relationship between tasks.

[0087] For each task T i , mark its dependent task set Dep(T i )={T j ∣T j →T i}, that is, T i Need to rely on T j It can be started only after execution is completed.

[0088] The execution order of tasks is calculated based on the task dependency graph, and tasks without dependencies (task nodes with in-degree 0) are scheduled first.

[0089] The priority calculation formula is:

[0090]

[0091] Determine the priority of task scheduling and dynamically migrate tasks to nodes with loads below a preset threshold based on the load status and resource availability of edge nodes;

[0092] Among them, T i represents the task, P(T i ) represents task T i priority, α1, β1, γ1 represent weight coefficients, T delay Represents task T i Delay tolerance, D dep Represents task T i The dependency depth, R res Represents task T i The amount of resources required, R avail Indicates the current available resources of the task execution node.

[0093] Construct edge node set N={N1,N2,…,N m}, each node N j Resource status R j ={C j ,M j ,B j}, representing computing power, memory capacity, and bandwidth respectively.

[0094] Using the distributed task allocation algorithm, the tasks to be executed are assigned to the nodes with lower load. By real-time monitoring of the resource utilization of each node U j , calculate the target node for task migration:

[0095]

[0096] Among them, U j For node N j The current load.

[0097] Dynamic task migration:

[0098] For resource-constrained nodes, dynamically migrate their tasks to the target node N target The migration process includes saving the task status, data transmission and restarting the task on the target node.

[0099] The scheduling of task migration follows the task dependency graph to ensure that the execution order of tasks after migration is not affected.

[0100] Dynamically adjust node resource allocation strategies based on node load status and task execution requirements. For high-priority tasks, more computing, memory, and bandwidth resources are allocated, while low-priority tasks are delayed or migrated to other nodes.

[0101] Define the node resource allocation formula:

[0102]

[0103] Among them, R total (N j ) is node N j The total available resources, R alloc (N j ) represents the edge node N j Assigned to task T i The amount of resources, T k Represents the kth task in the task set T.

[0104] A distributed task scheduling management module is established, where each node regularly reports its resource status and task execution progress to the scheduling center. Based on this feedback, the scheduling center dynamically updates the task dependency graph and node allocation strategy, ensuring that the scheduling logic is synchronized with resource status in real time.

[0105] The dynamic adjustment of task allocation and resource allocation strategies includes monitoring vehicle energy consumption status by expressing it in the form of a percentage of battery capacity, and dynamically updating the energy consumption rate based on the vehicle operating status (such as driving speed and task load).

[0106] Normalize the monitored data to generate the vehicle energy consumption state vector E = {E remain ,R consume} , as the input of scheduling decision, where R consume Indicates the energy consumption rate of the vehicle.

[0107] According to the resource requirements of the task R res and the estimated execution time T exec , calculate the energy consumption requirement E of each task task :

[0108] E task =R res ·T exec

[0109] Energy consumption demand is further divided into computing energy consumption E compute , communication energy consumption E comm and storage energy consumption E storage :

[0110] E task =E compute+E comm +E storage

[0111] For communication tasks, the communication energy consumption is calculated based on the transmission distance d and the amount of transmitted data D:

[0112] E comm =k·d α ·D

[0113] Where k is the transmission coefficient and α is the path loss exponent.

[0114] Prioritize tasks according to energy consumption requirements, prioritize low-energy consumption tasks, and prevent high-energy consumption tasks from occupying limited remaining power resources.

[0115] The priority calculation formula is:

[0116]

[0117] And adjust the computing resource allocation capacity of the edge node according to the remaining power status of the vehicle, and dynamically allocate it to R alloc Multi-level allocation strategy;

[0118] Among them, P task Indicates the energy consumption priority of the task, E remain Indicates the remaining power of the vehicle, E task Indicates the energy consumption requirement of the task. task Value, select tasks for scheduling from high to low.

[0119] Dynamically adjust the computing power of edge devices, optimize resource allocation based on the vehicle's remaining power and the task's energy consumption requirements. The computing power of edge devices is divided into multiple priority levels, each corresponding to a certain energy consumption threshold:

[0120]

[0121] Among them, θ high and θ low is the power threshold value, R max ,R mid ,R min To allocate computing power under different power conditions, R alloc Indicates the amount of computing resources allocated to edge nodes.

[0122] During the task execution, the actual energy consumption E of the task is monitored in real time. actual and estimated energy consumption E task Deviation ΔE=E actual -E task and update the energy consumption calculation model through online learning mechanism.

[0123] When the remaining power of the vehicle drops to the set threshold, the low-energy consumption mode is triggered, which restricts the scheduling of high-energy consumption tasks and prioritizes offloading tasks to other edge nodes.

[0124] For high-energy consumption tasks, edge nodes with sufficient resources are prioritized for task offloading based on the vehicle's energy consumption status and the load of edge devices.

[0125] The uninstall selection formula is:

[0126]

[0127] Among them, N target Uninstall the node for the target, E remain,j Represents node N j remaining available power.

[0128] The semantic analysis technology includes using a neural network model to perform semantic segmentation on the vehicle's surrounding environment data. The model uses an attention mechanism to improve the recognition ability of scene types such as traffic accidents, obstacles, or construction areas, and adjusts the task priority based on the scene type. The adjustment formula is:

[0129] P task,i =P base,i +δ·R(l i )

[0130] Among them, l i is the scene label, R(l i ) is scene l i The weight adjustment value for task priority, P task,i represents the adjusted priority of task i, P base,i represents the basic priority of task i, and δ represents the scenario impact factor.

[0131] The shortest path algorithm selects the optimal path for task transmission, including caching data required for key tasks at edge nodes, reducing access to cloud resources through cache management, and selecting cached data based on task priority, access frequency and dependencies.

[0132] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0133] Embodiment 2, the second embodiment of the present invention, is different from the previous embodiment in that:

[0134] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0135] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0136] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0137] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0138] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0139] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A real-time task scheduling method for cloud-edge collaboration in intelligent connected vehicles, characterized by: include, Collect real-time vehicle status data, mission requirements, and network conditions, and perform multi-source data fusion processing on vehicle sensors, V2X communications, and historical mission records to generate comprehensive data; Based on comprehensive data, the dynamic task priority optimization mechanism is used to adjust task priorities, and the shortest path algorithm is used to select the optimal path for task transmission based on network topology information, real-time network status, and energy consumption status; By allocating tasks collaboratively between edge nodes, dynamic task migration and load balancing are achieved. Task allocation and resource allocation strategies are dynamically adjusted based on the vehicle's remaining battery life and task energy consumption requirements. Semantic analysis technology is used to identify the scene type around the vehicle and dynamically adjust the task scheduling strategy. Execute tasks according to the optimized scheduling plan, monitor task execution status in real time, and update scheduling logic; The collaborative task allocation between edge nodes includes constructing a task dependency graph to represent the execution relationship and order between tasks, and calculating the priority using the following formula: Determine the priority of task scheduling and dynamically migrate tasks to nodes with loads below a preset threshold based on the load status and resource availability of edge nodes; Among them, T i represents the task, P(T i ) represents task T i priority, α1, β1, γ1 represent weight coefficients, T delay Represents task T i Delay tolerance, D dep Represents task T i The dependency depth, R res Represents task T i The amount of resources required, R avail Indicates the current available resources of the task execution node; The dynamic adjustment of task allocation and resource allocation strategy includes prioritizing low-energy consumption tasks by calculating the current remaining power of the vehicle and the energy consumption requirements of the tasks. The priority calculation formula is: And adjust the computing resource allocation capacity of the edge node according to the remaining power status of the vehicle, and dynamically allocate it to R alloc Multi-level allocation strategy; Among them, P task Indicates the energy consumption priority of the task, E remain Indicates the remaining power of the vehicle, E task Indicates the energy consumption requirement of the task; The semantic analysis technology includes using a neural network model to perform semantic segmentation on the vehicle's surrounding environment data. The model uses an attention mechanism to improve the ability to identify scene types such as traffic accidents, obstacles, or construction areas, and adjusts task priorities based on the scene type. The adjustment formula is: P task,i =P base,i +δ·R(l i ) Among them, l i is the scene label, R(l i ) is scene l i The weight adjustment value for task priority, P task,i represents the adjusted priority of task i, P base,i represents the basic priority of task i, and δ represents the scenario impact factor.

2. The real-time task scheduling method for cloud-edge collaboration in intelligent connected vehicles according to claim 1, characterized in that: The multi-source data fusion processing includes extracting temporal and spatial features from sensor data, V2X communication data, and mission record data through a deep learning model. Temporal features are extracted by a bidirectional long short-term memory network, and spatial features are extracted by constructing dynamic relationships between data nodes through a graph neural network combined with an attention mechanism to generate a comprehensive data set for scheduling decisions.

3. The real-time task scheduling method for cloud-edge collaboration for intelligent connected vehicles according to claim 2, characterized in that: The dynamic task priority optimization mechanism includes dynamically adjusting task priorities using a reinforcement learning algorithm, adjusting task priority weights based on the delay tolerance, resource requirements, and dependencies of the tasks by defining a state set, an action set, and a reward function, and updating the task scheduling order based on real-time changes.

4. The real-time task scheduling method for cloud-edge collaboration for intelligent connected vehicles according to claim 3, characterized in that: The method of selecting the optimal path for task transmission using the shortest path algorithm includes calculating the path weight based on network topology information and real-time network status, combined with delay, reliability, and energy consumption constraints, and selecting the task transmission path using the shortest path algorithm while verifying whether the delay, reliability, and energy consumption constraints of the path meet the scheduling requirements. The weight formula is: w ij =α·d ij -β·ln(r ij )+γ·c ij Among them, w ij Represents edge e ij The comprehensive weight of α, β, and γ represent weight coefficients, and d ij represents the communication delay between nodes i and j, r ij represents the communication reliability between nodes i and j, c ij represents the communication energy consumption between nodes i and j.

5. The real-time task scheduling method for cloud-edge collaboration for intelligent connected vehicles according to claim 4, characterized in that: The shortest path algorithm selects the optimal path for task transmission, including caching data required for key tasks at edge nodes, reducing access to cloud resources through cache management, and selecting cached data based on task priority, access frequency and dependencies.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images